Decalin ring structures in the products of boron fluoride-catalyzed

Decalin ring structures in the products of boron fluoride-catalyzed oligomerization of cyclopentene. Anatoli Onopchenko, Barrett L. Cupples, and A. No...
0 downloads 0 Views 633KB Size
Macromolecules 1982,15, 1201-1205

1201

Table I Polymerization of 2-MOCB with Modified (i-Bu),Al Initiatorsu added per mole of (i-Bu ), A1 0.5 H,Od 0.5 H,O 0.67 H,O 0.67 H,O 0.67 H,Oe 1.0 H,O 0.5 H,O, 0.5 Acac 0.5 H,O, 0.5 Acac 0.5 H,O, 0.5 Acac 0.4 H,O, 0.8 Acac 0.5 H,O, 1.0 Acac 1.0 H,O, 0.5 Acac

temp, "C

time, h

0

45

-18 60 -30 -30 -30 30

120 49

0 -30 30 30 30

0.25 88 25 266 1107 138 84 3 721 83 6

polymer yield, w t %

Gnbx

92 67 94 45 72 86 41 0.9 0.8 60 37 1.9

G w bx

[ri 1IC

dL/g

12 210 4 33 74 57 40

38 475 16 147 185 14 5 770

0.35 3.7

290 300

895 970

6.9

2.6

Polymerizations carried o u t in n-heptane solution under the following conditions (unless otherwise indicated): conDetermined by gel centration of monomer, 3.4 mol L - ' ; concentration of initiator ((i-Bu),Al basis), 0.07 mol L-'. Intrinsic viscosity was measured in chloroform a t 25 "C. Concentration of permeation chromatography in THF. monomer, 7.5 mol L - ' ; concentration of initiator, 0.15 mol L-'. e Concentration of monomer, 1.0 mol L - ' ;concentration of initiator, 0.02 mol L-'.

with the previous assignment of the signa1s.l It is noted that the triplet signals at 72.6, 72.7, and 72.8 ppm arising from the methine carbons in the stereochemical triads in the head-to-tail structures appear in intensity ratios very close to 1:2:1, indicating randomness of the configuration and no stereoselection by the applied initiator. The spectra l b and ICof the polymers prepared with the coordination initiators clearly indicate the absence of head-to-head, tail-to-tail structures as found also by Oguni and Hyoda for the polymer prepared with their initiator. Interestingly, we find, however, a marked shift in the intensity distribution of the triplet signals, appearing at the same positions as in spectrum l a from the triad structures, being 1:1:1.2 in the case of l b and 1:1.2:1.6 in the case of IC. Therefore, the distribution of the chiral carbons is not random in the polymers prepared with the coordination initiators. Thus these initiators are able to exert an influence on the stereochemistry of the propagation. I t is seen that the signal at 72.8 ppm arising from the syndiotactic triad, as already assigned to the lowest field,2grows in intensity at the expense of the signals from the heterotactic and isotactic triads. This effect appears to be accentuated by increasing the amount of acetylacetone in the initiator. However, as already pointed out, this amount must be limited. In the case of the cationic polymerization of 2-MOCB, stereospecificity is not found because both of the oxygen-carbon bonds in the cyclic oxonium ion may open in the propagation step, and it appears that the chain end is not able to exert a controling influence during the propagation of this monomer. Vandenberg found5y6with this type of cationic initiator that tactic polymers could be obtained in the case of the disubstituted epoxides cisand trans-2,3-epoxybutane,since with these monomers a chain end interaction with the oxonium ion causes the ring opening to take place in a specific manner or an enantiomeric selection takes place during the process. in the case of the coordination polymerization of 2-MOCB the basic condition for obtaining head-to-tail polymers is fulfilled, namely, that the bond breakage in the cyclic propagating intermediate occurs in a specific manner. The coordination complex also has some ability to select between the two enantiomeric forms of the monomer, yielding a less than random proportion of heterotactic structures. The nature of this controlling influence, however, has not been investigated, and it is not known presently if the polymers produced are homogeneous with respect to stereochemical distribution. 0024-9297/82/2215-1201$01.25/0

Experimental Section Materials. The monomer 2-MOCB was prepared as previously described.' Triisobutylaluminum (Fluka, practical grade) was used as received. n-Heptane was distilled and was more than 99.5% pure as determined by gas chromatography. Acetylacetone (Merck, synthesis grade) was distilled prior to use. Initiator Preparation. The preparations were carried out in a glovebox by adding water to a solution of ( ~ - B U ) in ~ An-~ heptane a t 0 "C under magnetic stirring. In the case of the preparation of the coordination initiator, acetylacetone was subsequently added after 30 min (0 "C). The solutions were stirred for 24 h a t room temperature prior to use. Polymerization Experiments. All polymerizations were carried out by using standard high-vacuum techniques. The initiator solutions were added with a syringe to the polymerization ampules. The polymerizations were terminated by addition of ethanol containing concentrated hydrochloric acid. The solutions were subsequently washed with 1N sodium hydroxide and water. Characterization of Products. Molecular weights were determined by gel permeation chromatography (Waters Model 200) using T H F solvent and polystyrene gel columns with porosities 1@,2 x lo4, 104,and lo3 A. The molecular weight values were calculated on the basis of a calibration curve obtained with narrow molecular weight samples of polystyrene from Water Associates. The calculations were based on the principle of extended chain lengths applying a Q value (weight per A) of 15.4. The "C NMR spectra were obtained in deuteriochloroform a t room temperature with tetramethylsilane as internal reference on a Bruker HX270 F T spectrometer operating a t 67.89 MHz.

References and Notes (1) Kops, J.; Hvilsted, S.; Spanggaard, H. Macromolecules 1980, 13, 1058. (2) Oguni, N.; Hyoda, J. Macromolecules 1980,13, 1687.

(3) Vandenberg, E. J. J.Polym. Sci., Part A-1 1969,7,525. (4) Kops, J.; Spanggaard, H. Polym. Bull. 1981,4, 505. (5) Vandenberg, E.J. Pure Appl. Chem. 1976,48, 295. (6) Ho, F. F.-L.; Vandenberg, E. J. Macromolecules 1979,12,212.

Decalin Ring Structures in the Products of BF&atalyzed

O l i g o m e r i z a t i o n of C y c l o p e n t e n e

ANATOLI ONOPCHENKO,* BARRETT L. CUPPLES, and A. NORMAN KRESGE Chemicals and Minerals Division, Gulf Research and Development Company, Pittsburgh, Pennsylvania 15230. Received November 23, 1981

Cationic polymerizations of alkenes are commercially important processes for the preparation of widely used 0 1982 American Chemical Society

1202 Notes

Macromolecules Scheme I Reaction Pathway for Oligomerization of Cyclopentene

f\r\

Q O+-W

-H +

-I4

II

+

w

2 0 Ni

f

111 ( 0 % )

+

w,. I".% -H

+

REAR..

Ye"+

ISOMERS

2%Ni V

03 00 -0 -H+

"2

~

t

VI1

NI

Vlll (TRANS. 33%: CIS, 67%)

"2

IX

polymers. Two excellent texts are available on the subject.'P2 While much information is available on the products of the polymerization of simple alkenes such as propylene and isobutylene, relatively little is known about the structure of products derived from cyclics. We were interested in the oligomerization of cyclopentene and cyclohexene with BF3 complexes. The literature on this subject was very limited. For example, depending on the alkene and the catalyst, the alkenes polymerize either (1)across the double bond, (2) with ring opening to give cis and/or trans polymers, or (3) both. Boor et al.3 reported that Ziegler catalyst (TiC1, and AEh) and a cationic vanadium system (VC4 and AlMe2Cl or AlMeJ both polymerized cyclopentene, while A1C1, was ineffective. The authors concluded that cyclopentene formed polymers resulting from both the polymerization across the double bond and ring-opening processes. None of the catalysts studied were effective with cyclohexene. In contrast, Rudenko and Gromova4 reported that the polymerization of cyclopentene with AlC13 (20% AlC13, 50 OC, 30 h) afforded the corresponding oligomers. The dimer fraction was identified largely as the saturated spiro[4.5]decane. Hofmann5 mentioned in his progress report that he polymerized both cyclopentene and cyclohexene with BF3 (HF promoter) and obtained dimers, trimers, tetramers, and still higher distillable oligomers, but the experimental conditions and analytical data were not given. Natta et a1.6 treated cyclopentene with cationic WC16 and obtained only a small amount of oil after 24 h of reaction. The same catalyst in the presence of A1Et3 converted cyclopentene into essentially pure trans-polypentenamer. Finally, Nametkin and Rudenko' treated cyclohexene with AlCl, (510% AlCl,, 70-80 "C, 30 h) and found among the oligomers a considerable amount of saturation. One

of the hydrocarbons was identified as tetracyclohexylbenzene.

Discussion Recently, Cesca et al.svg reported on the structure of products resulting from the cationic polymerization of propylene and 1-butene. Using 13CNMR, they identified 11 distinct structural groups, but no predominant isomerization pathway was evident. This finding suggested to us that during the oligomerization of cyclic alkenes, some isomerization in the way of ring contraction and ring expansion should be present in addition to the usual reactions across the double bond and ring opening, but this isomerization has not been demonstrated. In this study, cyclopentene and cyclohexene were oligomerized with BF3, with 1-butanol as the cocatalyst. The products were oligomeric, with the degree of polymerization n being -2-6. Analysis of the original products by 'H and 13C NMR indicated the absence of aromatic structures and products from ring-opening reactions. Hence, the cyclic structures were preserved in the BF, system. To simplify the analysis, the reaction mixture was hydrogenated, and the analysis of the saturated hydrocarbon mixture thus obtained was performed by GLC after the product had been separated into its respective oligomer fractions by distillation. In this manner, the products from the ring expansion, those from ring contraction, and those from the unrearranged skeleton could most readily be established. The hydrogenated dimer fraction from the cyclopentene reaction was found to consist of trans-decalin (22%), bicyclopentyl(33%), and cis-decalin (44%). The results of oligomerization of cyclopentene are explained according to Scheme I. Following the protonation of cyclopentene, the cyclopentyl carbenium ion adds to a second cyclopentene to give ion I. Ion I can rearrange to ion 11, but

Vol. 15, NO.4,July-August 1982

Notes 1203

Scheme I1 Reaction Pathway for Isomerization of 3-Cyclopentylcyclopentene

“^“a) -m+a IV

-& REAR.

IV

-H

-26%

VI

VI1

-H+l -74%

80 H2

I

+

Ni

H2

I

03 Vlll

(TRANS, 68%: CIS, 42%)

00 IX

this rearrangement did not appear to take place since none of the bicyclo[5.3.0]decane (111)was detected among the products. Ion I can undergo deprotonation to form alkenes, but apparently it first rearranges to a more stable ion IV, which eventually will give bicyclopentyl (V). Ion IV apparently also rearranges to the spiro[4.5]decyl ion VI, which then further rearranges to VI1 and results in the decalin ring structure. Ion IV on deprotonation was expected to give some spiro[4.5]dec-l-ene, but none of this alkene was detected. In a control experiment 3-cyclopentylcyclopentene,when treated with BF3under the reaction conditions, was isomerized to a mixture of octalins (26%) and spiro[4.5]dec1-ene (74%), which on hydrogenation resulted in cis-decalin (11% ), tram-decalin (15%), and spiro[4.5]decane (74%), supporting the proposed rearrangement mechanism (Scheme 11). The basic difference between the products of cyclopentene oligomerization (Scheme I) and those of 3-cyclopentylcyclopentene isomerization (Scheme 11)is the predominance of the spirodecane ring formation in the latter with only little oligomerization. This difference can be explained on the basis of steric hindrance. For example, ions VI and VII, derived from 3-cyclopentylcyclopentene, are sterically hindered and are considered unpolymerizable. Hence, the only reaction that can occur is deprotonation to form alkenes. In the presence of cyclopentene, however, the alkenes formed from ions VI and VI1 can react further with cyclopentyl carbenium ions to form oligomers via a series of proton elimination/alkylation reactions. In this case, the spiro[4.5]dec-l-ene formed would represent the least hindered alkene and therefore was expected to react further. Of some interest is a paper by Naro and Dixon,lo who studied the alumina and sulfuric acid catalyzed dehydration of 6-hydroxyspiro[4.5]decane. Dehydration over alumina at elevated temperature produced approximately equal amounts of alkenes having the decalin and the spiro[4.5]decane ring structures. Dehydration by sulfuric acid, on the other hand, produced mostly the decalin ring. With neither catalyst was the expected bicyclopentyl de-

tected. It would appear that the type of product formed is determined by the kinetic vs. thermodynamic control of the reaction. Other studies on the solvolysis of neopentyl-type tosylates of the spiran series show that a five-membered ring was readily converted to a six-membered ring via ring expansion, while in the case of a sixmembered ring, the course of reaction was dependent on the size of the ring bearing the leaving group.’lJ2 The trimers of cyclopentene could result from the addition of cyclopentyl carbenium ions to alkenes derived from ions IV,VI, and VII. This fraction, however, proved to be too complex for detailed analysis, although the predominance of decalin ring structures was evident from the 13C NMR data.13 Our observations with cyclopentene in the BF, system show that about one-third of the expected reaction occurred and that the major course of reaction at the dimer stage (two-thirds) involved rearrangement. This rearrangement is unique, as the cyclopentylcyclopentenium ion IV undergoes two successive ring expansions to give the decalin ring structure. The formation of decalin rings via ring expansion is not new, but the formation of these from cyclopentenehas never been observed prior to the present study. Hence, in any future work with cyclopentene, particularly when cationic catalysts are used, researchers should look for the presence of decalin ring structures among the products of the reaction. The dimer fraction from cyclohexene oligomerization, by analogy to the cyclopentene oligomerization, was expected to consist of bicyclohexyl and the cis- and transbicyclo[5.5.0]dodecanes. Analysis of this fraction by GLC showed three components in 1,4, and 95%,respectively, in order of their emergence from the column. When the dimer fraction was spiked with bicyclohexyl, the largest peak in the chromatogram corresponded with the bicyclohexyl peak. The identity of bicyclohexyl was also confirmed by mass spectrometry (MS/GLC)14 and 13C NMR. Mass spectrometry showed the two minor compounds to be isomeric Cl2Hz2( m / e 166) but did not reveal their identities. Comparison of the fragmentation patterns of these minor compounds with those of the published

1204 Notes

Macromolecules

spectra for model compounds ruled out the presence of spiro[5.6]dodecane and dimethyldecahydronaphthalene structure^.'^ Since the mass spectral data for cis- and trans-bicyclo[5.5.0]dodecaneswere not available, these isomers were prepared by the lithium amide catalyzed isomerization of 1,5,9-~yclododecatrienefollowed b y hydr0genati0n.l~ Chromatographic analysis showed that bicyclo[5.5.0]dodecanes were not present i n t h e dimer fraction. The most diagnostic clue to the identity of the predominant minor component came from the 13C NMR spectrum, which showed a distinct peak at 15.6 ppm for the methyl group. The compound i n question is believed to be l-methyl-l-cyclohexylcyclopentane,formed via ring contraction of cyclohexyl carbenium ion t o l-methylcyclopentenium ion followed by addition to cyclohexene. Such a possibility was supported by the fact that the recovered cyclohexene contained some methylcyclopentenes. In a related work, Laber16dehydrated 7-hydroxyspiro[5.6]dodecane using zinc chloride and obtained an 80% yield of cyclohexylcyclohexene. In solvolysis studies of the same spiro system, Krapcho and co-workers1' identified their products, after hydrogenation, as spiro[5.6]dodecane (15% ), bicyclohexyl(80% ), and an unidentified component (5%). It appears therefore that the formation of two fused seven-membered rings via rings enlargement is energetically unfavorable. In contrast t o cyclopentene, the oligomerization of cyclohexene proceeds mostly in the expected manner, but with some rearrangement involving ring contraction (-5%). The BF, oligomerizations require a promoter and an excess of BF, over that needed to form a 1:l molar complex with the promoter. The degree of oligomerization i n the BF, reaction can be varied by changes in promoter concentration or type and in reaction t e m p e r a t ~ e . 'The ~ most common promoters include water, alcohols, and carboxylic acids. The unique advantage of m a n y BF,.HA complexes as catalysts apparently lies in the moderate acid strength of the equilibrium BF3.HA + H+BF3A-.18 Competing complexation and decomplexation reactions such as BF,*HA HA + BF3.2HA and BF,*HA + BF3 HA (A = n-BuO) explain the necessity of keeping the system saturated with BF,. Quenching of the reaction mixture with caustic liberates the promoter, which is then washed o u t with water.

+

+

Experimental Section Both oligomerization and hydrogenation experiments were carried out in a l-L, 316 stainless steel, magnetically stirred autoclave that was equipped with a stirrer, a heating mantle, and cooling coils (The Autoclave Engineers, Inc.). In a typical experiment, 300 g of cyclopentene (or cyclohexene), 170 g of nhexane, and 3 g of l-butanol were charged into the autoclave and pressured with BF3 to 100 psig. The temperature was adjusted to 3 0 4 0 "C, the stirring rate was set at lo00 rpm, and the reaction was arbitrarily continued for 1 h, periodically adding BF, to maintain a pressure of 100 psig. At the end of the run, the product was discharged into 500 mL of 25% sodium hydroxide solution to give a black, tarlike material, which, when shaken, turned into a bright yellow product. The organic layer was separated and washed 4 times with 400-mL portions of water (final wash was neutral). The product was dried (anhydrous MgS04),filtered over Celite, and hydrogenated over Ni 0104 catalyst (10 g, 200 "C, loo0 psig of H2, 1 h). After stripping of cyclopentane and n-hexane at atmospheric pressure, the residue (158 g) was distilled via a glass-packed, five theoretical plate, Oldershaw column to give oligomeric fractions. The distribution of oligomers was determined on a Hewlett-Packard 5710A gas chromatograph using an l/g in. X 2.5 ft, 5% Dexsil300 on Chrom W packed column, programed from 90 to 390 "C a t 32 "C/min (thermal conductivity detector temperature, 350 "C; injection port temperature, 400 "C; attenuation 1). Detailed composition of isomers was determined with

a 0.01 in. x 150 ft poly(pheny1 ether) (OS-138) column at 100 "C. The distribution of oligomers from cyclopentene at 54% cyclopentene conversion was as follows: dimer, 39.2% ; trimer, 35.2%; tetramer, 15.3%; pentamer, 10.2%. The dimer fraction, bp 47-55 "C (3.5 mmHg), was found by GLC to consist of trans-decalin (22%), bicyclopentyl(33%), and cis-decalin (44%), identified by comparison of their retention times to those of authentic standards. 13CNMR (CsD6,Me4%): trans-decalin, 6 27.24 (2, 3,6, 7), 34.72 (1,4, 5, 8), 44.22 (9, 10); cis-decalin, 6 24.59 (2, 3, 6, 7), 29.77 (1,4, 5,8), 36.87 (9, 10); bicyclopentyl, 6 25.80 (3,4,3', 4'), 32.24 (2,5, 2', 5'), 46.89 (1, 1'). The integration was in agreement with the GLC data. Analysis of the hydrogenated trimer fraction (GLC),bp 116-143 "C (0.65 mmHg), showed two major components (16 and 51%, respectively) and eight minor components (33%), ranging in composition from 1 to 9%. The mass spectrum (MS/GLC) showed the major compounds to be isomeric C15H26 ( m / e 206), but no structural information was deduced. The 13C NMR spectrum showed a predominance of decalin ring structures (major peaks at 6 21.96, 25.57, 26.41,27.78, and 29.02; intermediate peaks at 6 23.97, 36.62,38.33,39.55, and 42.65; minor peaks at 6 23.29, 24.60, 29.57, 29.88, 31.05, 32.08, 37.44, 41.89, 45.55, 47.17, and 49.26). The oligomerization of cyclohexene was carried out under the exact conditions used with cyclopentene. The distribution of oligomers at 74% conversion was as follows: dimer, 44.6%; trimer, 14.5%; tetramer, 35.6%; pentamer, 5.0%. The hydrogenated dimer fraction, bp 63-68 "C (0.6 mmHg), consisted of bicyclohexyl (95%) and two minor isomers (1 and 4%, respectively). The 13C NMR spectrum was identical with that of authentic bicyclohexyl (6 27.34 (3, 4, 5 , 3', 4', 5'), 30.65 (2, 6, 2', 6'), and 43.88 (1,1'))but additionally contained weak peaks at 6 15.60 (CH,) and 6 21.59, 21.91,25.90,33.66,36.20,37.70, and 39.52. One of the minor compounds is believed to be 1methyl-l-cyclohexylcyclopentane. The hydrogenated trimer fraction, bp 138-156 "C (0.6 mmHg), too complex for detailed analysis, showed three major peaks by GLC (27, 10, and 26%, respectively) and 17 minor peaks, ranging in composition from 1 to 6% each. The 13C NMR spectrum showed an intense peak a t 6 27.30, moderate peaks a t 6 30.42, 30.72, and 43.85, and weak peaks at 6 21.97, 22.22,32.30,34.24, 37.94, 38.55, and 39.75. Control Experiment. A mixture of 3-cyclopentylcyclopentene (40 g), cyclohexane (250 g), and l-butanol (0.4 g) was reacted in the autoclave (30 "C, 50 psig of BFd for 10 min to trap the initially formed intermediates. On workup as before, followed by distillation, a monomer fraction, bp 40-55 "C (2-3 mmHg), was obtained that contained 1- and 3-cyclopentylcyclopentenes(82%), A19-octalin(2%), Ag(")-octalin (3%), and spiro[4.5]decenes (13%). M e r hydrogenation (PtO,, EtOH, 25 "C, 50 psig of H2),cis-decalin ( l l % )trans-decalin , ( E % ) , and spiro[4.5]decane (74%) were detected in the product (bicyclopentyl-free basis), with little oligomers. Synthesis of Bicyclo[5.5.0]dodecanes. trans,trans,cis1,5,9-Cyclododecatriene (57 g) was added to a solution of the lithium derivative of ethylenediamine (25 g of Li in 530 mL of amine), and the mixture was stirred a t 95-100 "C for 1h. After hydrolysis, extraction with ether, and distillation, 29.9 g of a mixture of bicyclo [ 5.5.01dodeca-1,7-diene and bicyclo [ 5.5.01dode~a-A'(~)-2-diene was obtained bp 61-63 "C (0.4 mmHg) [lit.13 bp 97.5-99.5 "C (8 mmHg)]; 'H NMR (CC14,Me4Si) 6 0.7-2.5 (m, 16 H), 5.2-5.9 (m, 2 H). Hydrogenation of the diene mixture over Ni 0104 (200 "C, lo00 psig of H,) followed by distillation afforded 13.5 g of bicyclo[5.5.0]dodecane: bp 53-57 "C (0.4 mmHg) [lit.13bp 106-107.5 "C (9 mmHg)]; 'H NMR 6 1.1-1.9 (m); GLC, two isomers (14 and 86%); mass spectrum (MS/GLC) of the first isomer (probably trans), m / e 166.2 (11%) (M+),82.1 (loo), 96.2 (83), 67.1 (61), 81.1 (57);mass spectrum of second isomer (cis), m / e 166.2 (20%) (M'), 96.2 (loo), 81.1 (97), 82.1 (91), 67.1 (73).

References and Notes (1) Plesch, P. H. "The Chemistry of Cationic Polymerizations";

Macmillan: New York, 1963.

(2) Kennedy, J. P. "Cationic Polymerization of Olefins: A Critical Inventory"; Wiley: New York, 1975.

Macromolecules 1982,15, 1205-1207 Boor, J.; Youngman, E. A.; Dimbet, M. Makromol. Chem. 1966,90,26. Rudenko, M. G.; Gromova, V. N. Dokl. Akad. Nauk SSSR 1949,67,885. Hofmann, F. Chem. Ztg. 1933,57,5. Natta, G.;Dall'Asta, G.; Mazzanti, G. Angew. Chem., Znt. Ed. Engl. 1964,3,723. Nametkin, S.S.;Rudenko, M. G. J . Gen. Chem. USSR (Engl. Transl.) 1937,7,763. Ferraris, G.; Corno, C.; Priola, A,; Cesca, S. Macromolecules 1980,13,1104. Priola, A.; Corno, C.; Cesca, S. Macromolecules 1980,13,1110. Naro, P. A.; Dixon, J. A. J . Am. Chem. SOC.1959,81, 1681. Krapcho, A. P.; McCullough, 3. E.; Vahabedian, K. V. J. Org. Chem. 1965,30,139. Krapcho, A. P.; Benzon, M. J . Am. Chem. SOC.1962,84,1035. Dalling, D.K.;Grant, D. K.; Paul, E. G. J. Am. Chem. SOC. 1973,95,3718. Cornu, A.; Massot, R. "Compilation of Mass Spectral Data", 2nd ed.; Heyden: New York, 1966,1975;Vol. 1. Crain, D. L.; Hartzfeld, H. A. US. Patent 3009001;Chem. Abstr. 1961,56,8594d. Laber, G . Justus Liebigs Ann. Chem. 1954,588,79. Brennan, J. A. Znd. Eng. Chem. Prod. Res. Dev. 1980,19,2. Shubkin, R. L.; Baylerian, M. S.; Maler, A. R. Znd. Eng. Chem. Prod. Res. Dev. 1980,19,15.

Monte Carlo Calculation of the Hydrodynamic Radius at the 8 Point. Deviations from Analytical Gaussian Behavior CHARLES M. GUTTMAN,* FRANK L. MCCRACKIN, and

1205

what has been proposed before. We compute the diffusion coefficient from the Kirkwood theory with preaveraging for polymer chains at the 8 point by calculating the 1/R moment of these chains created by a Monte Carlo simulation in which both excluded volume and nearest-neighbor interaction are included.'O The value of p obtained for such chains is closer to the experimentally obtained p and differs from that obtained for an analytical Gaussian coil. The value of p obtained from this Monte Carlo calculation, pMC, is 1.40 f 0.01 for chains with an infinite number of monomers. 11. Gaussian Coil Inverse Radius: Slow Approach to Asymptotic Value In Kirkwood' theory, the diffusion constant, D , is given by (3)

where the first term is the free-draining limit, which depends on the friction coefficient ({) and the number of statistical segments in the chain (N). Substituting eq 3 in eq 2 yields 1 (4) RH where r

CHARLES C. HAN

National Measurement Laboratory, National Bureau of Standards, Washington, D.C. 20234. Received September 21, 1981

I. Introduction A number of recent papers have been published comparing theoretical'9 expressions for the diffusion coefficient for a polymer at infinite dilution with the diffusion coefficient obtained from quasi-elastic light ~ c a t t e r i n g . ~ ~ In a recent comprehensive comparison of theory and experiment, Schmidt and Burchard3 found that there is a discrepancy between the predicted and theoretical values of the dimensionless parameter pE PE = RG/RH (1) where RGis the z-average radius of gyration of a polymer chain, usually obtained from light scattering or neutron scattering experiment$, and RH is the experimentally obtained mean reciprocal hydrodynamic radius defined3in terms of the Stokes law expression

where vo is the solvent viscosity, kT the thermal energy, and D the measured z-average translational diffusion coefficient for the polymer obtained from quasi-elastic light scattering. Schmidt and Burchard3found pE to be 1.27 f 0.05 for a large number of experiments on polystyrene in 8 solvents. The above pE deviates from the 1.504 theoretical value of p predicted by Kirkwood theory' for the analytical Gaussian coil model of the polymer. These deviations of experiment and theory have yet to be explained, although usually the inadequacies of the Kirkwood dynamical formulation are considered to be the cause of the discrepancy.2+' In this paper, we approach the observed deviation of theory and experiment from a point of view different from

The value of h* has been estimated in a variety of ways for various models of the chain. For chain models where b'and N refer to quantities defined in the spring-and-bead model of a polymer chain, Wangl' considers values of h* in the range 0.12-0.40. Because the term involving h*, the free-draining term, is considered small compared to the ( l / R ) term for large N, Schmidt and Burchard3 did not consider it in their estimation of l/RH. For the moment, we, too, shall not consider this term further but rather focus on the calculation of ( l / R ) . The term ( l / R ) in Kirkwood theory is defined as

where rij is the distance between beads i and j and n is the number of bonds in the chain. For the analytical Gaussian coil model one has12

The Gaussian (1/R) was obtained from eq 6 and 7 by replacing the sum by its trapezoid rule integral. The approximate formula

to order 1/N was obtained. Equation 8 gives the term in 1/Nf2 exactly but gives the coefficient of the 1/N only approximately. Although inexact, the above result shows the slow approach to the asymptotic value of the Gaussian ( l / R ) , (l/R)G, to its limiting value. The mean reciprocal inverse radius has also been computed for values of N to 100000 using eq 6 and 7 by performing the summation on a com-

This article not subject to US. Copyright. Published 1982 by the American Chemical Society